Optics Communications 285 (2012) 4949–4953

Contents lists available at SciVerse ScienceDirect

Optics Communications

0030-40

http://d

n Corr

E-m

journal homepage: www.elsevier.com/locate/optcom

Modeling and testing of static pressure within an optical fiber cable spoolusing distributed fiber Bragg gratings

Chengju Ma a,b, Liyong Ren a,n, Enshi Qu a, Feng Tang c, Quan Liang c

a State Key Laboratory of Transient Optics and Photonics, Xi’an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Xi’an 710119, Chinab School of Science, Xi’an Shiyou University, Xi’an 710065, Chinac Xi’an Institute of Modern Control Technology, Xi’an 710049, China

a r t i c l e i n f o

Article history:

Received 27 October 2011

Received in revised form

20 July 2012

Accepted 21 July 2012Available online 4 August 2012

Keywords:

Fiber cable spool

Fiber optic guided missile (FOG-M)

Fiber Bragg grating (FBG)

Axial direction strain

Radial direction strain

Pressure

18/$ - see front matter & 2012 Elsevier B.V. A

x.doi.org/10.1016/j.optcom.2012.07.064

esponding author. Tel./fax: þ86 29 88881538

ail address: renliy@opt.ac.cn (L. Ren).

a b s t r a c t

Based on the force analysis, we establish a theoretical model to study the static pressure distribution of

the fiber cable spool for the fiber optic guided missile (FOG-M). Simulations indicate that for each fiber

layer in the fiber cable spool, the applied static pressure on it asymptotically converges as the number

of fiber layers increases. Using the distributed fiber Bragg grating (FBG) sensing technique, the static

pressure of fiber cable layers in the spool on the cable winding device was measured. Experiments

show that the Bragg wavelength of FBG in every layer varies very quickly at the beginning and then

becomes gently as the subsequent fiber cable was twisted onto the spool layer by layer. Theoretical

simulations agree qualitatively with experimental results. This technology provides us a real-time

method to monitor the pressure within the fiber cable layer during the cable winding process.

& 2012 Elsevier B.V. All rights reserved.

1. Introduction

Since the fiber cable is impervious to conventional electroniccounter-measures, it emits no exploitable radio frequency signals,provides extremely wide bandwidth and a high data transmissionrate, it offers a major benefit for the fiber optic guided missile(FOG-M) technology [1]. In the FOG-M system, the fiber cable isfirstly wound on a spool, then stored for later use, and finally releasedfrom the spool to server as the bidirectional communications mediumfor transmitting both the situational data from the missile to thelaunch platform and the command guidance signals from the launchplatform to the missile. Obviously, the winding and releasing techni-ques of the fiber cable are two key technologies in designing a reliableFOG-M system. Note that the pressure in the spool affects the fiberbirefringence and thus the signal transmission quality, therefore theinner pressure analysis of the FOG-M fiber cable spool is not onlynecessary for designing the fiber cable winding and releasingmechanism, but also significant for analyzing the loss of the fibercable and for predicting the lifetime of the fiber cable spool.

Since the fiber Bragg grating (FBG) has lots of advantages suchas its small size, anti-interference ability, high sensitivity, goodstability, long lifetime, wide monitoring range, real-time onlinemeasurement, easy to form smart sensor network and other

ll rights reserved.

.

advantages [2–6], it is widely used in sensing field [7–10]. FBGacts as an optical filter that reflects the incident light at aparticular wavelength (Bragg wavelength), and this reflectionwavelength changes in proportion to the applied strain ortemperature change [11,12].

In this paper, we propose a new method to study the innerpressure distribution in the fiber cable spool. By introducing thedistributed FBGs to the spool, we realize a real-time measurementto the fiber cable layer pressure by use of the FBG sensingcharacteristic of the radial direction strain. A theoretical modelof the fiber cable layer’s pressure distribution for the fiber cablespool was established. Theoretical simulations agree qualitativelywith experimental results. Using this method, the problem of thefiber cable layer pressure real-time monitor during the cablewinding process was solved. This research provides a goodreference to the fiber cable winding, the fiber cable layer pressuremeasurement, the spool design, the fiber cable loss analysis andthe lifetime prediction of the fiber cable spool in the FOG-Msystem. Moreover, this study might provide a valuable referenceto the fiber winding in the fiber optic gyroscope [13] and to thefilament winding in the composite pressure vessels [14], etc.

2. Theoretical model

The fiber cable spool of FOG-M is processed by using a specialwinding device to wind the fiber cable on the spool. However,

C. Ma et al. / Optics Communications 285 (2012) 4949–49534950

we can simplify the complicated winding device to a simplemechanism model, as shown in Fig. 1, for the convenience to thetheoretical analysis. A constant tension force T to the fiber isprovided by the weight G. The fiber is wound to the spool layer bylayer as the spool is accurately rotated and moved up and down.To establish a theoretical model of the fiber cable layer’s pressure,we depict Fig. 2 to analyze the layer’s force. In what follows, wegradually analyze this theoretical model from the simple condi-tion to the close actual situation. Firstly, we analyze the simplestcase where both the radial shrink of the fiber cable spool and thefriction force between fiber cable layers are neglected. In thiscase, we suppose that the fiber cables align regularly along thespool axial direction and have no deformation. Secondly, weconsider the case where the radial shrink of the fiber cable spoolis involved. As in the actual spool of the FOG-M system, thepressure from the outer cable layers causes a slight displacementto the inner fiber cables, which generally makes the cablesbecome tighter and experience a radial shrink. Finally, the frictionforce between fiber cable layers is further considered. Throughthese three steps, the ideal theory model is revised gradually andbecomes suitable to the actual fiber cable spool of FOG-M.

2.1. Simplest case

Fig. 2 illustrates the force analysis of the fiber cable spoolsupposing that only the first fiber cable layer is wound on it. R isthe spool radius, O is axis center of the spool, dl is the fiber cablelength micro-element, dy is the plane angle corresponding to dl,and N is the crutch force. Thus, the stress to the spool surface canbe written as

F0 ¼�N¼ 2T sindy2: ð1Þ

Consider that the relation of sin(dy/2)Edy/2¼dl/2R is alwayssatisfied since dy is very small for the micro-element dl, one caneasily obtain the pressure putting on the spool surface given by

P0 ¼F0

Ds¼

F0

Ddl¼

1

D

T

R, ð2Þ

where D is the fiber cable diameter and Ds¼Ddl is the micro-element area.

G

T

Fig. 1. Schematic of the fiber cable winding device, where weight G provides

a constant tension force T to the fiber.

TT

θd

dl

R

O x

TT

N

y

2/θd

Fig. 2. Force analysis of the fiber cable spool on which just one layer of fiber cable

was wound.

Based on the above analysis, the tension force exerted on everyfiber cable is same as T when one neglects the spool strain and thefriction force between fiber cable layers. In this case, through asimple analysis and deduction, one can express the pressureexerted on the m-th fiber cable layer, Pm, for the total layer onthe spool, n, in the following form:

Pm ¼1

D

Xn�1

i ¼ m

T

Rþ iD0, ð3Þ

where D0 ¼31/2�D/2 is the distance between the middle of two

layers.

2.2. Considering radial direction shrink of fiber cable spool

In practical situation, the pressure from outer fiber cable layerscan cause the inner cables align more tightly and deformation. So,the distance between cable layers will decrease. Then, a shrink ofthe cable spool along the radial direction to its center will happen.Obviously, such a shrink may inevitably lead to a change in theaxial direction tension force and the radial direction pressure ofthe inner cable. Let us take the m-th cable layer as an example,the radial shrink strain can be expressed as

emr ¼DR

RþmD0¼ �

Pm

ar, ð4Þ

where ar is the radial direction shrinking coefficient of the spool,DR is the radial direction shrink of the m-th cable layer. Note thatthis radial direction shrinking may result in the changes in theaxial length, the axial tension force and the radial directionpressure of the fiber cable. Let us explain these three changesone by one. Firstly, the axial shrink of the fiber cable in the m-thlayer, Dlm, denoted by the fiber length change within one windingcircle, can be calculated as

Dlm ¼ 2pDR¼�2pRþmD0

arPm: ð5Þ

Secondly, the axial tension force of the fiber cable in the m-thlayer, denoted by Tm, is accordingly changed from the constant T

to TþDTm. It should be pointed out that, for the radial shrink case,DTm is negative in our simulations. And it can be obtainedthrough the relation [15]

DTm

S¼ E

Dlmlm

, ð6Þ

where E and S are the axial elasticity modulus and the crosscutting area of the fiber cable, respectively, and lm is the one-winding-circle perimeter of the fiber cable in the m-th layer asgiven by

lm ¼ 2pðRþmD0Þ: ð7Þ

Inserting Eqs. (5) and (7) into Eq. (6), one can obtain the axialactual tension force of the fiber cable in the m-th layer in thefollowing form:

Tm ¼TþDTm ¼ T�

ES

arPm ðm¼ 1,2,. . .,n�1Þ

T ðm¼ nÞ:

8<: ð8Þ

Taking into account that different layer has different tensionforce due to the radial shrink of the spool (as given by Eq. (8)), onecan re-write Eq. (3) as

Pm ¼1

D

Xn�1

i ¼ m

Tiþ1

Rþ iD0: ð9Þ

Using Eqs. (8) and (9) and gradually iterating from the outerlayer (the n-th layer, Pn¼0) to the inner layer, one can achieve,after a series of interminable but not difficult deductions, the

500 m fiber 50 m fiber

FBG1 FBG2

FBG10

FBG11

1000 m fiber450 m

Fig. 3. Diagram of the FBGs fused into the fiber cable.

C. Ma et al. / Optics Communications 285 (2012) 4949–4953 4951

radial pressure of the m-th fiber cable layer as

Pm ¼T

D

Xn�1

i ¼ m

1

Rþ iD0�

ES

ar

� �1

D

� �2

TXn�2

i ¼ m

1

Rþ iD0Xn�1

j ¼ iþ1

1

Rþ jD0

0@

1A

þES

ar

� �2 1

D

� �3

TXn�3

i ¼ m

1

Rþ iD0Xn�2

j ¼ iþ1

1

Rþ jD0Xn�1

k ¼ jþ1

1

RþkD0

0@

1A

0@

1A

�ES

ar

� �3 1

D

� �4

TXn�4

i ¼ m

1

Rþ iD0Xn�3

j ¼ iþ1

1

Rþ jD0Xn�2

k ¼ jþ1

1

RþkD0

0@

0@

�Xn�1

z ¼ kþ1

1

RþzD0

!!!þ � � � ¼

Xn�m

s ¼ 1

�1ð Þs�1 ES

ar

� �s�1 1

D

� �s

T

�Xn�s

i ¼ m

1

Rþ iD0Xn�sþ1

j ¼ iþ1

1

Rþ jD0Xn�sþ2

k ¼ jþ1

1

RþkD0Xn�sþ3

z ¼ kþ1

1

RþzD0

0@

0@

� � � �Xn�1

w ¼ qþ1

1

RþwD0

!� � �

!!!!: ð10Þ

We emphasize that Eq. (10) has already included the abovementioned third change, i.e., the radial direction pressure of the fibercable, resulted from the radial direction shrink effect. And, in whatfollows, we will indicate that Eq. (10) can be further simplified whenone considers the related parameters according to actual fiber cablespool system. Based on the actual situation, the parameters in Eq. (10)are listed here: T¼200 N, D¼340 mm, R¼0.05 m, ar¼1.30�1011 Pa,E¼7.452�1010 Pa. Based on the numerical relationship of RcD, it iseasy to find that from the third term in Eq. (10) each term can beregarded as an infinitesimal compared to its previous adjacency term.Therefore, we may approximate Eq. (10) by just keeping its first twoterms as given by

Pm ¼T

D

Xn�1

i ¼ m

1

Rþ iD0�

ES

ar

� �1

D

� �2

TXn�2

i ¼ m

1

Rþ iD0Xn�1

j ¼ iþ1

1

Rþ jD0

0@

1A: ð11Þ

Eq. (11) means that the radial pressure of fiber cable layer isquite insensitive to the radial shrink of the fiber cable spool.However, the axial shrink of the fiber cable originated from theradial shrink of the fiber cable spool will result in a sensitivechange in the FBG’s central wavelength. Note that we will discussthis point in Section 2.3 in detail.

2.3. Considering friction force between fiber cable layers

Considering the friction force between fiber cable layers, theaxial direction tension force of the fiber cable in the m-th layercan be given by

T 0 ¼ T�mPmpðRþmD0ÞD, ð12Þ

where m is the static friction coefficient. Replacing T in Eq. (11)with T0 expressed by Eq. (12), we thus deduce the radial pressureof the m-th fiber cable layer as given by

Pm ¼

TD

Pn�1i ¼ m

1Rþ iD0� ES

ar

� �1D

� �2TPn�2

i ¼ m1

Rþ iD0Pn�1

j ¼ iþ11

Rþ jD0

� �1þmpðRþmD0Þ

Pn�1i ¼ m

1Rþ iD0� ES

ar

� �1D

� �mpðRþmD0Þ

Pn�2i ¼ m

1Rþ iD0

Pn�1j ¼ iþ1

1Rþ jD0

� � :ð13Þ

In the experiment, in order to measure the radial pressuredistribution with the fiber cable layer, a serial of FBGs withdifferent Bragg wavelengths were fused into the fiber cable aspressure sensors. It is well-known that FBG’s Bragg (center)wavelength change resulted from the axial direction strain isgiven through the following relation [15]:

DlBz ¼ lBð1�peÞezm, ð14Þ

where lB is the FBG’s Bragg wavelength at free state, pe is the fibereffective elasto-optic coefficient, ezm¼Dlm/lm is the FBG’s axialdirection strain. Note that Dlm and lm have already defined byEqs. (5) and (7), respectively. The subscript ‘z’ indicates the axialdirection of the fiber cable. Inserting Eqs. (5) and (7) intoEq. (14) and considering the sign of Dlm, the change in Braggwavelength of FBG in the m-th fiber layer is then given by

DlBz ¼�lBð1�peÞ

arPm: ð15Þ

On the other hand, considering that the radial directionpressure Pm is exerting on the FBG, this may lead to an additionalshift to the FBG Bragg wavelength, as given by [16]

DlBr ¼�kprlBPm, ð16Þ

where kpr is the radial pressure sensitivity coefficient of FBG. Notehere we use the subscript ‘r’ to indicate the radial direction of thefiber cable. Obviously, under both the axial-direction and theradial-direction strains, the change in Bragg wavelength of FBGcan be calculated as

Dl¼DlBzþDlBr ¼�lB1�pe

arþkpr

� �Pm, ð17Þ

where ar is the radial direction shrinking coefficient. Although ar

is usually a constant, for our case it varies as the fiber winds onthe spool since the spatial gap between fiber cable layers alwayschanges during the winding process. Considering this effect,Eq. (17) should be revised if one takes into account the actualsituation of fiber winding process. The dependence can beexplained as follows. On one hand, with the increase of m, theradial direction shrinkage of the m-th fiber cable layer becomesgreater and greater. On the other hand, with the increase of thetotal layer number n, the inner fiber cable layer becomes increas-ingly tight and the axial direction shrink becomes increasinglysmaller. Taking into account this two factors, the shrinkagecoefficient, ar, needs a modification to satisfy the practicalsituation. Based on the above analysis, we hereby suppose thatthe radial shrink coefficient depends on m and n through ar/m�n.In this case, Eq. (17) should be modified and denoted as

Dl¼�lB1�pe

ðar=mÞ � nþkpr

� �Pm: ð18Þ

Eq (18) denotes the Bragg wavelength shift of the FBG in them-th fiber cable layer caused by both the axial- and the radial-direction strains, where the total layers wound on the fiber spoolis n. The comparison between the experimental results and thesimulations, showing in the following Sections 3 and 4, indicatethat such a modification to the radial shrink coefficient isbasically correct.

3. Experimental results

In order to measure the static pressure at different layers ofthe fiber cable spool, we need to embed the sensing elements(FBGs) into the fiber cable. Fig. 3 depicts the schematic diagram toshow how the FBGs are arranged along the fiber cable. Thiscorresponds to a fiber cable we fabricated and used in theexperiment, where 11 FBGs were fused into the fiber cable witha length of 2 km. The Bragg wavelengths of the 11 FBGs are from1534 nm to 1554 nm with a wavelength interval of 2 nm.

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

1110 98 765 43 2 1

Fiber cable layerFBG1 (1534nm)FBG2 (1536nm)FBG3 (1538nm)FBG4 (1540nm)FBG5 (1542nm)FBG6 (1544nm)FBG7 (1546nm)FBG8 (1548nm)FBG9 (1550nm)FBG10 (1552nm)FBG11 (1554nm)

FBG distributed positionon fiber cable spool (m)

Fibe

r ca

ble

laye

r nu

mbe

r n

Fig. 4. Diagram of the FBG’s positions in the fiber cable spool.

5 10 15 20 25 30 35 40

-0.8

-0.6

-0.4

-0.2

0.0

Cha

nge

of F

BG

Bra

ggw

avel

engt

h �

� (n

m)

Fiber cable layer number n

5 10 15 20 25 30 35 40 45 50

-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0m=7 (1534nm)m=8 (1536nm)m=9 (1538nm)m=10(1542nm)m=11(1544nm)m=12(1546nm)m=13(1548nm)m=14(1552nm)m=15(1554nm)

Cha

nge

of F

BG

Bra

ggw

avel

engt

h �

� (n

m)

Fiber cable layer number n

0 5 10 15 20 25 30 35 40 45-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0m=8 (1536nm) simulation

m=8 (1536nm) experiment

m=10(1542nm) simulation

m=10(1542nm) experiment

m=12(1546nm) simulation

m=12(1546nm) experiment

Cha

nge

of F

BG

Bra

gg w

avel

engt

h �

� (n

m)

Fiber cable layer number n

Fig. 5. Relationship between the FBG’s Bragg wavelength shift Dl and fiber cable

layer number n. (a) for the experiment, (b) for the simulation and (c) for the

comparison.

C. Ma et al. / Optics Communications 285 (2012) 4949–49534952

The fabrication process of the FBG-embeded fiber cable can besimply described as follows: firstly, the first FBG with thewavelength of 1534 nm was fused at 500 m location, as shownin Fig. 3, secondly, the other 10 FBGs with correspondingwavelengths were fused into the fiber with a distance intervalof 50 m and thirdly, the last 1000 m fiber was fused. Finally, sucha FBG-embeded fiber was re-coated and recovered to a fiber cablethrough a special treatment process.

Fig. 4 shows the schematic and relative positions of the elevenFBGs in the spool. Note that for the first layer the wound cablelength is totally 78 m where the cable is wrapped on the spoolwith a radius of 0.05 m. In order to ensure that every fiber cablelayer to be tightly wound on the spool thus to prevent the layercollapse, during the winding process, the two ends of each layerare gradually shrunk to the center. The black small square marks,with the corresponding number labels, indicate the positions ofthe 11 FBGs in the spool. For example, the first FBG is located onthe 7th layer when the wound cable length reaches 500 m.

We tested the whole winding process. When it reaches the 7thfiber cable layer (n¼7), the first FBG was wound on the spool.At this moment, the Bragg wavelengths of the FBGs string weredemodulated and the first FBG’s Bragg wavelength was obtained.And then, with a continuous winding, the second FBG was woundon the spool when the cable layer number was 8 (n¼8). At thismoment, the Bragg wavelengths of the FBG string were demodu-lated again and the first and the second FBG’s Bragg wavelengthswere obtained. We did in this way until the 2 km fiber cable wascompletely wound on the spool.

Fig. 5(a) shows the Bragg wavelengths variations of the elevenFBGs during the whole experimental winding process. We can seethat the FBG’s Bragg wavelengths in the fiber cable layers varyvery quickly at the beginning and then become gently. Thisfeature is consistent to the fact that the inner cable layers becomegradually tight and stable as the cable layers increase. Note that,as compared with other FBGs, the 4th FBG’s Bragg wavelengthvariation is distinct. This is because that the 4th FBG is locatednear the edge of the layer, it is influenced by fewer layers above it.(see Fig. 4).

4. Theoretical simulation

It is seen from Eq. (18) that the Bragg wavelength shift of FBGdepends on two factors. Firstly, the radial direction shrink of thefiber cable spool causes every layer FBG axial direction strain.

Secondly, the pressure from the outer fiber cable layers causes theradial direction strain of the FBG located in the inner layer.Using Eqs. (13) and (18) we simulated the Bragg wavelengthshifts of FBGs and the pressures for fiber cable layers from the 7thlayer to the 15th one, supposing that there are 50 layers beingwound on the spool. The Bragg wavelengths of FBGs are from1534 nm to 1554 nm with an interval of 2 nm. The parametersused in simulations are the actual experimental parameters. i.e.,Pe¼0.22, ar¼1.30�1011 Pa, kpr¼8.38�10�13 Pa�1, R¼0.05 m,D¼340 mm, T¼200 N, m¼0.00625.

0 5 10 15 20 25 30 35 40 45 50

0

50

100

150

200

m=7

m=8

m=9

m=10

m=11

m=12

m=13

m=14

m=15

Fibe

r ca

ble

laye

r pr

essu

re P

m (

MPa

)

Fiber cable layer number n

Fig. 6. Theoretical dependence of the fiber cable layer pressure Pm on the fiber

cable layer number n.

0 5 10 15 20 25 30 350

50

100

150

200m=7 (FBG1)m=8 (FBG2)m=9 (FBG3)m=9 (FBG4)m=10 (FBG5)m=11 (FBG6)m=12 (FBG7)m=13 (FBG8)m=13 (FBG9)m=14 (FBG10)m=15 (FBG11)

Fibe

r ca

ble

laye

r pr

essu

re P

m (M

Pa)

Fiber cable layer number n

Fig. 7. Variation of the fiber cable layer pressure Pm calculated according the

experimental data with the fiber cable winding layer n.

C. Ma et al. / Optics Communications 285 (2012) 4949–4953 4953

The theoretical variation of the FBG Bragg wavelength with theincrease of the fiber cable layer is shown in Fig. 5(b). It is seen thatthe Bragg wavelength of each FBG changes rapidly at the begin-ning and then tends to stability. In order to compare thetheoretical and the experimental results, we select three curvesfrom both Fig. 5(a) and (b) and plot them on Fig. 5(c). Althoughthe simulations agree qualitatively with the experiments, therestill exists a considerable difference between them. We currentlythink that this difference might be caused by two factors. Firstly,as compared to the constant tension force T in the simulation, itinevitably has a fluctuation during the process of fiber winding.Secondly, the modification of the shrinkage coefficient, ar, mightbe more complicated than that in Eq. (18).

Shown in Fig. 6 is the dependence of the fiber cable layerpressure Pm on the fiber cable layer number n. Here we select m

from 7 to 15 and n from 0 to 50 as example. It seems that Pm willgo to convergence as n further increases.

In fact, the winding process is a dynamic stabilizing process,i.e., the inner cables become gradually tight and stable owing tothe radial direction shrink of the spool with the increase of the

layers. Therefore the shifts of Bragg wavelengths and the varia-tions of the pressures become stable gradually.

Based on the experimental data of Bragg wavelengths varia-tion of the eleven FBGs shown in Fig. 5(a), using Eq. (18) weachieve the relationship between radial direction pressure Pm ofthe fiber cable layer and the fiber cable layer number n, as shownin Fig. 7. It is seen that Fig. 7 basically coincides with thestimulation result (see Fig. 6).

5. Conclusion

In conclusion, we established a theoretical model for describingthe inner static pressure among fiber cable layers in the spool ofthe FOG-M based on the force analysis of the system. Both thesimulations and the experiments were conducted. Simulationsindicate that the radial pressure of fiber cable layer in the spoolasymptotically converges as the number of fiber layers increase. Atthe same time, the FBG’s Bragg wavelength in the fiber cable layervaries very quickly at the beginning and then becomes gently. TheBragg wavelength variation of FBG is resulted from a combinedeffect including both the FBG axial strain caused by the spool radialshrink and the FBG radial strain caused by the pressure from outercable layers during the winding process. The simulation resultsagree qualitatively with the experimental ones. This technologyprovides us a real-time method to monitor the pressure of the fibercable layer during the cable winding process. This research has asignificant reference in many areas such as the fiber cable spoolwinding, the pressure measurement of the fiber cable spool, thefiber cable spool design.

Acknowledgments

This work was partially supported by the West Light Foundationof the Chinese Academy of Sciences (CAS) (2009LH01), the Innova-tion Foundation of CAS (CXJJ-11-M22), the Advanced Programs ofTechnological Activities for Overseas Scholars, and the ScientificResearch Foundation for the Returned Overseas Chinese Scholars,State Education Ministry.

References

[1] A.R. Habayeb, IEEE LCS 2 (1991) 18.[2] E.J. Friebele, C.G. Askins, M.A. Putnam, A.A. Fosha Jr., J. Florio Jr., R.P. Donti,

R.G. Blosser, Electronics Letters 30 (1994) 1783.[3] H.J. Sheng, W.F. Liu, K.R. Lin, S.S. Bor, M.Y. Fu, Optics Express 16 (2008) 16013.[4] Y.J. Rao, Optics and Lasers in Engineering 31 (1999) 297.[5] A. Arie, B. Lissak, M. Tur, Journal of Lightwave Technology 17 (1999) 1849.[6] X.W. Shu, L.Z. Yu, D.H. Zhao, L. Zhang, K. Sugden, I. Bennion, Journal of the

Optical Society of America B 19 (2002) 2770.[7] J.X. Dai, M.H. Yang, Y. Chen, K. Cao, H.S. Liao, P.C. Zhang, Optics Express

19 (2011) 6041.[8] H. Tsuda, Optics Letters 35 (2010) 2349.[9] D. Gatti, G. Galzerano, D. Janner, S. Longhi, P. Laporta1, Optics Express

16 (2008) 1945.[10] J.L. Gan, Y.Q. Hao, Q. Ye, Z.Q. Pan, H.W. Cai, R.H. Qu, Z.J. Fang, Optics Letters

36 (2011) 879.[11] W.W. Morey, G. Meltz, W.H. Glenn, Proceedings of SPIE 1169 (1989) 98.[12] A.D. Kersey, M.A. Davis, H.J. Patrick, M. LeBlanc, K.P. Koo, C.G. Askins,

M.A. Putnam, E.J. Friebele, Journal of Lightwave Technology 15 (1997) 1442.[13] J. Nayak, Applied Optics 50 (2011) E152.[14] V.V. Vasiliev, A.A. Krikanov, A.F. Razin, Composite Structures 62 (2003) 449.[15] W.T. Zhang, L.H. Liu, F. Li, Y.L. Liu, Chinese Optics Letters 5 (2007) 507.[16] H.J. Sheng, M.Y. Fu, T.C. Cheng, W.F. Liu, S.S. Bor, Photonics Technology

Letters 16 (2004) 1146.